HSDWT Based Multispectral and Hyperspectral Image Fusion

Multispectral (MS) and hyperspectral (HS) image fusion is a sub-branch with comparatively less research achievements in the field of image fusion. According to the characteristics and 3D feature analysis of the MS and HS image data volumes, a new fusion algorithm using Hyperspectral Discrete Wavelet Transform (HSDWT), HSDWT-MHF algorithm, is proposed. This algorithm is composed of four major steps: spatial and spectral resampling, HSDWT, wavelet coefficients integration and inverse HSDWT. The results of simulated experiments illustrate that compared with 2D wavelet transform based traditional fusion approach, the newly proposed HSDWT based fusion algorithm is capable of preserving both spatial and spectral features of the source images more adequately and producing fused image with higher quality and fewer artifacts.